The sample points must be finite.

I like to interpolate a sample data on a new timeline by useing interp1, saying "The sample points must be finite."

data_origin=readtable('data_origin.csv');

time=data_origin.Var1;

elev=data_origin.Var2;

%% note the start date in raw data is {'05-Jan-2020 17:40:00'}

%% end date in raw data is {'09-Nov-2020 16:20:00'}

timeline = ( datetime(2020,1,5) + minutes((1060:25:24*60*310-5*92)) ).' ;

%% interpolate schism result to 'obs_tide_time'

interp_elev = interp1(datenum(time(:)),double(elev(:)),datenum(timeline(:)),'linear','extrap');

Answer 1

[Askic V]

Most likely, you're getting the error because you have missing entries in the data.

websave("filename.csv","https://www.mathworks.com/matlabcentral/answers/uploaded_files/1415214/data_origin.csv");

data_origin = readtable("filename.csv");
timeData = data_origin.Var1;
values = data_origin.Var2;

% Convert datetime to numerical values
timestamps = datenum(timeData);

% Remove NaN or Inf values from the arrays
validIndices = isfinite(timestamps);
timestamps = timestamps(validIndices);
values = values(validIndices);
data_origin(188:190,:)

ans = 3×2 table
            Var1            Var2
    ____________________    ____

    08-Jan-2020 23:35:00    0.44
                     NaT    0.47
    09-Jan-2020 00:25:00     0.5

% Is there missing entries in the data:
missing_entries = any(isnat(data_origin.Var1) | isnan(data_origin.Var2));
if missing_entries
    % remove mising entries
    data_origin = rmmissing(data_origin);
end

% Define new extrapolation range
extrap_start = datenum('05-Jan-2020 17:10:00');
extrap_end = datenum('10-Nov-2020 00:20:00');

% Convert new timestamps to numerical values
extrap_timestamps = linspace(extrap_start, extrap_end, numel(values));

% Perform linear extrapolation using interp1
extrap_values = interp1(timestamps, values, extrap_timestamps, 'linear', 'extrap');

% Convert numerical timestamps back to datetime format
extrap_datetime = datetime(extrap_timestamps, 'ConvertFrom', 'datenum');

plot(extrap_datetime, extrap_values)

[jie hu]

Really appreciated, Askic!